Background Survival estimates are integral to care for patients diagnosed with dementia. Few Canadian studies have carried out long-term follow-up of well-described cohorts, analyzing survival related to multiple risk factors. Methods Survival analysis of an inception cohort enrolled at a British Columbia (BC) tertiary dementia referral clinic between 1997 and 1999 was undertaken. Vital status was completed for 168 patients diagnosed with dementia. An evaluation of the effects of demographics, vascular risk factors, cognitive and functional ratings, apolipoprotein 4-status, and cholinesterase use on survival was performed using a log-rank test and time-dependent Cox regression. Survival of this dementia cohort was compared with the age-matched life expectancy of persons in BC. Results In all, 158/168 (94.0%) subjects died over 16.6 years, with a median survival of 7.08 years. Risk factors associated with shorter survival in dementia groups included age of onset ≥80 (hazard ratio [HR] 1.56, 95% confidence interval [CI] 1.05-2.32); greater functional disability (Disability Assessment for Dementia<55% [HR 1.47, 95% CI 1.04-2.08]); and cumulative medical illness severity (Cumulative Illness Rating Scale≥7 [HR 1.51, 95% CI 1.08-2.12)]. Compared with the BC population, years of potential life lost for dementia subjects aged <65 was 15.36 years, and for dementia subjects aged ≥80 it was 1.82 years. Conclusions Survival in dementia subjects is shorter than population life expectancies for each age strata, with greatest impact on younger patients. For people diagnosed with dementia, age ≥80 years, cumulative medical illness severity, and functional disabilities are the most significant survival predictors and can be used to guide prognosis.